CN104980353A

CN104980353A - Ethernet virtual private tree (e-tree) service with optimal forwarding in ethernet virtual private network (evpn)

Info

Publication number: CN104980353A
Application number: CN201510146514.2A
Authority: CN
Inventors: 萨米·布索斯; 阿里·萨嘉西; 萨姆尔·萨拉姆
Original assignee: Cisco Technology Inc
Current assignee: Cisco Technology Inc
Priority date: 2014-04-01
Filing date: 2015-03-31
Publication date: 2015-10-14
Anticipated expiration: 2035-03-31
Also published as: US20170310598A1; US20150281096A1; US9742678B2; EP2928114B1; CN104980353B; US10270697B2; EP2928114A1

Abstract

The invention discloses an ETHERNET VIRTUAL PRIVATE TREE (E-TREE) SERVICE WITH OPTIMAL FORWARDING IN ETHERNET VIRTUAL PRIVATE NETWORK (EVPN). In one embodiment, when an ingress provider edge (PE) device of a computer network domain receives a frame at the ingress PE device destined to a destination media access control (MAC) address, it can determine whether the frame was received on a root or leaf Ethernet ingress segment, and also whether the destination MAC address is located via a root or leaf Ethernet segment. Accordingly, the ingress PE device may either drop or forward the frame based on the ingress Ethernet segment and destination MAC address Ethernet segment being either a root or a leaf, respectively.

Description

There is in EVPN the optimum E-TREE service forwarded

Technical field

Disclosure relate generally to computer network, relates more specifically to have the optimum E-TREE service forwarded in EVPN.

Background technology

Ethernet virtual private tree (Ethernet Virtual Private Tree, or have point-to-multipoint Ethernet virtual link (the Ethernet Virtual Connection of root E-TREE), EVC) be by Metro Ethernet Forum (Metro-Ethernet Forum, MEF) the 2nd layer of service defined, the 2nd layer of service provides Ethernet VLAN (VLAN) configuration being applicable to multicast service.Illustratively, Internet Engineering Task group (Internet Engineering TaskForce, the IETF) internet draft of " the Requirements for MEF E-Tree Support in VPLS " <draft-ietf-12vpn-etree-reqt> by name proposed by people such as Key defines supports the requirement that MEF E-TREE serves in layer 2 virtual private network (L2VPN).Other types EVC for carrier ethernet net definitions is E-Line and E-LAN.

Website in E-TREE service has limited connectivity, and is designated as root and/or leaf.This service is established as follows:

-Gen website can communicate with other websites all (root or leaf).

-leaf site can communicate with root website, but can not communicate with other leaf site.

In L2VPN (be Virtual Private LAN Service (VPLS), virtual private multicast service (VPMS) or Ethernet Virtual Private Network (EVPN)), egress filtering model is all depended on to all schemes that E-TREE carries out addressing.This means outlet (that is, disposing) provider edge (PE) equipment determines it is forward or abandon the flow going to local connecting circuit, to meet the restriction of E-TREE connectivity.This model unnecessarily wastes the bandwidth of multiprotocol label switching (MPLS) network, many destinations flow (broadcast, unknown clean culture and multicast (BUM) flow) that its middle period copies to leaf flow, all known unicast traffics and entrance is transmitted in MPLS network, is only dropped in outlet PE.

Accompanying drawing explanation

Reference description below can understand embodiment here better by reference to the accompanying drawings, and reference number identical in accompanying drawing represents equivalent or intimate element, wherein:

Fig. 1 shows exemplary computer network;

Fig. 2 shows example network device/node;

Fig. 3 shows the exemplary view of the computer network for root segment and leaf section;

Fig. 4 shows example form;

Fig. 5 A-5C shows the example of unicast frame ingress filtering;

Fig. 6 shows the example that many destinations frame entrance forwards; And

Fig. 7 shows in the EVPN according to one or more embodiment described herein the exemplary reduced process with the optimum E-TREE service forwarded.

Embodiment

general view

According to one or more embodiment of the present disclosure, when entrance provider edge (PE) equipment in computer network territory receives at entry PE equipment place the frame going to destination medium access control (MAC) address, it can judge that this frame is received in root Ethernet port section or leaf Ethernet port section, and can judge that destination MAC Address is located via root ethernet segment or leaf ethernet segment.Therefore, entry PE equipment can be root or leaf based on entrance ethernet segment and destination MAC Address ethernet segment respectively, abandons or transmitted frame.

describe

Computer network is the set geographically distributed of node by the communication link for transmitting data between the end node of such as personal computer and work station and so on and section interconnection.The very eurypalynous network comprising local area network (LAN) (LAN) and wide area network (WAN) is all available.LAN is usually by being positioned at the dedicated private communications links connected node in same, physical (such as, building or campus).On the other hand, WAN connects by long-distance communication links (such as, common carrier telephone wire, optical beam path footpath, Synchronous Optical Network (SONET) or synchronous digital hierarchy (SDH) link) node geographically disperseed usually.The Internet is the example of the WAN connecting whole world heterogeneous networks, provides global communication between the node on various network.Node is usually by exchanging discrete frame or packet at the enterprising Serial Communication of network according to predefined agreement (such as, transmission control protocol/Internet protocol (TCP/IP)).Under this background, protocol package contains one group of how mutually mutual rule of defined node.Computer network can interconnect further by the intermediate network node of such as router and so on, to expand effective " size " of each network.

Because the management of the computer network to interconnection can prove heavy, so the group of less computer network can be maintained as routed domain or autonomous system.Network in autonomous system (AS) usually by tradition " in the territory " router couples being configured to Routing Protocol in execution domains to together, and usually by same mechanism control.In order to improve route extensibility, AS can be divided into multiple " region " or " rank " by service provider (such as, ISP).But may expect to increase can the quantity of node of swap data; In this case, perform the inter domain router of inter-domain routing protocol to be used to interconnect the node of various AS.In addition, also may expect that the various AS to operating under different administrative domains interconnect.As used herein, AS, region or rank are collectively referred to as in " territory ".

Fig. 1 is the schematic block diagram of exemplary computer network 100, and as shown, exemplary computer network 100 comprises the node/equipment by such as multiple router/equipment of link or network interconnection and so on illustratively.Such as, customer edge (CE) router one 10 can with provider edge (PE) router one 20 (such as, PE-1, PE-2 and PE-3) interconnection, to be communicated by the core net of such as illustrative multiprotocol label switching (MPLS) core net 130 and so on.Predefined network communication protocol can be used (such as, transmission control protocol/Internet protocol (TCP/IP), User Datagram Protoco (UDP) (UDP), asynchronous transfer mode (ATM) agreement, frame relay protocol or other suitable agreement any) by the node/equipment room exchange data packets 140 (such as, flow/message) of link at computer network 100.It will be understood by those skilled in the art that the node, equipment, link etc. that any amount in computer network, can be used, and shown here view is for simplicity.Those skilled in the art also will understand, although embodiment described herein is described for operating between AS usually, the disclosure can be applied to network configuration between any similar territory, and wherein similar technology will be suitable.

Fig. 2 is the schematic block diagram of example endpoint/equipment 200, and this example endpoint/equipment 200 may be used for one or more embodiment described herein, such as, as any router (especially PE 120) shown in Fig. 1.Equipment 200 comprises the multiple network interfaces 210, one or more processor 220 and the memory 240 that are interconnected by system bus 250.That network interface 210 comprises machinery, electric signal circuit, for transmitting data on the physical link being coupled to network 100.Network interface can be configured to use various different communication protocol to send and/or receive data.Note, physical network interface 210 can also be used to realize one or more virtual network interface well known by persons skilled in the art (such as, for virtual network interface that Virtual Private Network (VPN) is accessed).

Memory 240 comprises can by multiple memory locations of one or more processor 220 and network interface 210 addressing, for storing the software program and data structure that are associated with embodiment described herein.Processor 220 can comprise the necessary element or logic that are suitable for software program for execution and manipulation data structure 245.Operating system 242 (such as, Cisco Systems Inc. network interconnection operating system or ) multiple parts usually to reside in memory 240 and to be performed by one or more processor, it comes functionally to organize node by the network operation calling software process and/or the service that the equipment of being supported in performs.These software process and/or service can comprise route service 244 and illustrative ingress filtering process 248 is (as described herein, ingress filtering process 248 (such as, process " 248a ") can be positioned at independent network interface alternatively).

It will be apparent for a person skilled in the art that other processor and type of memory (comprising various computer-readable medium) can be used to store and perform the program command relevant with technology described herein.In addition, although specification describes various process, clearly can predict various process and can be embodied in and be configured to carry out according to technology (such as, according to the function of similar processes) herein the module that operates.In addition, although process is shown separately and/or describes, the process of it will be appreciated by those skilled in the art that can be other in-process module or routine.

Routing daemon/service 244 comprises the computer executable instructions performed by processor 220, to perform by one or more Routing Protocol (such as, Interior Gateway Protocol (IGP) (such as, Open Shortest Path First " OSPF " and intermediate system to intermediate system " IS-IS "), Border Gateway Protocol (BGP) etc.) function that provides, as being readily appreciated by one skilled in the art.These functions can be configured to manage forwarding information database, and this forwarding information database comprises such as making the data forwarding and determine.Particularly, the change in network topology can use the Routing Protocol of such as traditional OSPF and IS-IS link-state protocol and so on to be transmitted among routers 200 (such as, with " gathering " the same view to network topology).

Note, route service 244 can also perform the function relevant to virtual routing protocols, such as, safeguard the agreement of VRF instance or tunnel transmission such as MPLS, general MPLS (GMPLS) etc., as being readily appreciated by one skilled in the art.In addition, such as, the EVPN described in the IETF internet draft of " BGP MPLSBased Ethernet VPN " <draft-ietf-12vpn-evpn> by name introduces the scheme of serving for multiple spot L2VPN, the program has advanced many host ability, uses BGP dispatch user/client medium access control (MAC) address on core MPLS/IP network to arrive ability information.

As mentioned above, Ethernet virtual private is set (E-TREE) or is had the point-to-multipoint Ethernet virtual link (EVC) of root to be the 2nd layer of service defined by Metro Ethernet Forum (MEF), and the 2nd layer of service provides Ethernet VLAN (VLAN) configuration being applicable to multicast service.Website in E-TREE service has limited connectivity, and is designated as root and/or leaf.This service is set up as follows: root website can communicate with other websites all (root or leaf), and leaf site can carry out communicating with root website but can not communicate with other leaf site.Such as, as shown in Figure 3, according to the collocation strategy that it will be appreciated by those skilled in the art that, some ethernet segment (connecting circuit) can be marked as root (" ROOT ") or leaf (" LEAF ").

But, as mentioned above, for L2VPN (no matter be Virtual Private LAN Service (VPLS), virtual private multicast service (VPMS) or EVPN) in egress filtering model is all depended on to all schemes that E-TREE carries out addressing.This means outlet (that is, disposing) PE determines it is forward or abandon the flow going to local connecting circuit, to meet the restriction of E-TREE connectivity.This model unnecessarily wastes the bandwidth of MPLS network, many destinations flow (broadcast, unknown clean culture and multicast (BUM) flow) that its middle period copies to leaf flow, all known unicast traffics and entrance is transmitted in MPLS network, is only dropped in outlet PE.

On the other hand, technology herein provides the mechanism supported for the E-TREE in EVPN, wherein, utilizes entrance to copy and carries out ingress filtering optimization to clean culture and many destinations flow.Particularly, according to the of the present disclosure one or more embodiment be described below in detail, the entry PE equipment in computer network territory learns new MAC Address in this locality, and judges that MAC Address is all learnt on root ethernet segment or leaf ethernet segment respectively.Like this, when receiving at entry PE equipment place the frame going to destination MAC Address, it can judge that this frame is received in root entry section or leaf entrance, and can judge that destination MAC Address is located via root ethernet segment or leaf ethernet segment.Therefore, entry PE equipment can based on entrance and destination MAC Address section is respectively root or leaf abandons or transmitted frame.

Illustratively, technology described herein can by hardware, software and/or firmware according to such as, ingress filtering process 248/248a and routing daemon 244 perform, wherein, ingress filtering process can comprise by processor 220 (or independent processor of interface 210) execution to perform the computer executable instructions of the function relevant to technology described herein.Such as, technology herein can be regarded as the extension to conventional edge router protocol, and therefore can correspondingly be processed by the similar assembly of these agreements of execution understood in this area.

Operationally, for the ingress filtering of unicast frame, when on connected ethernet segment, this locality learns new MAC Address to EVPN PE, its this address of instruction in MAC advertising line (advertisement route) is learnt on root segment or leaf section.This can be realized illustratively by the mark introducing instruction root/leaf instruction in EVPN MAC advertising line or in new bgp attribute.Fig. 4 shows simplification table 400 (such as, transmit, data structure 245), this simplification table can store M AC address 405, corresponding forwarding information 410 and the new root/leaf instruction introduced.It will be appreciated by those skilled in the art that, simplification table 400 be only for illustration of, and do not mean that the scope of restriction this paper embodiment (such as, out of Memory and/or the form that can realize analog result also can be used by according to technology herein, and root segment or leaf section are learnt).

Therefore, according to technology herein, when PE receives MAC advertising line, in it is transmitted, " root/leaf " instruction is set for each MAC Address.This can be the unit marking be associated with each mac address table entry.Therefore, when PE receives the ethernet frame going to given MAC Address, following process will be performed:

-judge that this frame comes up at root segment or leaf section; And

-perform mac address table to destination-address to search, to judge that destination corresponds to root or leaf.

Then, entry PE can forward rule (that is, root takes office any station point, any website communicates to leaf to root with without leaf) according to E-TREE, and the combination of searching based on the root/leaf appointment and MAC that enter interface decides to abandon or transmitted frame.

Fig. 5 A-5C shows some examples of the unicast frame ingress filtering according to technology herein.More specifically, Fig. 5 A shows such example: frame (such as, grouping 140, represents with thick arrow) is received by from root segment at entry PE-1 place, and goes to the leaf section more than PE-2.Because this provenance-destination pairing is Gen-Ye, so this frame is allowed to.(note, when frame is received from root segment, the decision that the forwarding based on ingress filtering still abandons grouping can be simplified, because root segment can communicate with root or leaf, the further process that therefore destination root/leaf is specified is unnecessary).In addition, in figure 5b, Ye-Gen communication is also allowed to, and wherein PE-1 is from leaf section received frame, and frame goes to root segment.On the contrary, as shown in Figure 5 C, when receiving from leaf section the frame going to another leaf section, PE-1 abandoned (ingress filtering) this frame before this frame enters core net towards PE-2.

Except unicast frame, technology herein also can be applied to the ingress filtering of many destinations frame.Such as, a kind of like this mechanism is defined in the IETF internet draft of " E-TREE Support in EVPN " the <draft-sajassi-12vpn-evpn-etreeGreatT.GreaT .GT by name proposed by people such as Sajassi, can on the basis of each ethernet segment by this machine-processed PE, (A-D) route is to announce leaf/root instruction to use EVPN Ethernet automatically to find.According to one or more embodiments herein, also define a kind of mechanism, when entrance copies, the PE receiving those advertisements can come to perform ingress filtering to many destinations frame by this machine-processed usage data.

Particularly, the illustrative mechanism of this paper is as follows: on the basis of each EVPN example (EVI), PE checks all Ethernet A-D routes received from given long-range " PEr ".According to those routes, local PE can judge PEr whether have the root segment of connection, leaf section or they both.If PEr has both root segment and leaf section (such as, PE-2), then as the root segment that this PEr has for process subsequently, treat this PEr.On the contrary, if it only has leaf section (such as, PE-3), then it is regarded as leaf PE.

When constructing each bridge domain (BD) inundation list, the mark that PE indicates root/leaf to indicate marks the inundation list of entries of each long-range PEr, as above (such as, in the mode similar to the table in Fig. 4) that judge.When many destinations frame is received by PE on this location, it is received on root segment or leaf section that PE marks this frame.If this location is leaf section, then PE iteration and do not copy any frame to the PE being marked as leaf in inundation list of entries.

Such as, as shown in Figure 6, PE-1 can receive the frame of the multiple destinations gone to more than PE-2 and PE-3 from leaf section.Because PE-2 is listed in root PE (that is, it has one or more root segment), so frame is copied (forwarding) to PE-2 by PE-1.But therefore PE-1 will to PE-3 (ingress filtering), because PE-3 only has leaf section, and only not have leaf will terminate to the communication of leaf by duplicated frame.

Note, this mechanism can not be eliminated and perform the demand (such as, PE-2) of filtering for outlet PE.On the contrary, it is mainly paid close attention to solve and frame is sent to the problem of the PE abandoning them (such as, PE-3) being wasted broadband.More specifically, as in figure 6, the frame being sent to PE-2 will need egress filtering, be sent out in any leaf section correspondingly to prevent this frame.

Fig. 7 shows in the EVPN according to one or more embodiment described herein the exemplary reduced process 700 with the optimum E-TREE service forwarded, especially from the visual angle of entry PE (wherein, computer network territory can be E-TREE, root wherein communicates with leaf with other roots all, and leaf only communicates with root).Process 700 can start in step 705 place, and proceeds to step 710, and in step 720, as more described in detail above, PE learns new MAC Address in this locality.In a step 715, PE judges that MAC Address all learns on root ethernet segment or leaf ethernet segment (such as, based on the mark in MAC advertising line or based on bgp attribute field) respectively.In step 720, PE can transmit in (such as, table 400) for the instruction that each MAC Address section of setting is root segment or leaf section.Note, for long-range PE equipment (outlet PE), such PE have root segment, leaf section or they both, and wherein specific outlet PE equipment has both root segment and leaf section, and specific outlet PE is marked as root.

When receiving at entry PE equipment place the frame going to destination MAC Address in step 725, entry PE judges that this frame is received in root entry section or leaf entrance in step 730.In addition, in step 735, entry PE judges that destination MAC Address is located via root ethernet segment or leaf ethernet segment.According to root and leaf intercommunication rule, entry PE can by based on entrance and destination MAC Address section being root or leaf abandons or transmitted frame performs ingress filtering respectively in step 740.That is, as above, website in E-TREE service has limited connectivity, and wherein root website can communicate with other websites all (root or leaf), and leaf site can carry out communicating with root website but can not communicate with other leaf site.Note, in step 740, when frame goes to multiple destination, when frame is received in leaf section, frame can be forwarded to any outlet PE equipment being marked as leaf by prevention and perform many destinations ingress filtering by entry PE.

Then simplification process 700 can terminate in step 745, it should be noted that it is optional for learning new MAC Address in step 720 and in step 725, receive further frame.Although it should be noted that some step in process 700 is optional as mentioned above, the step shown in Fig. 7 be only for illustration of example, and some other step can be included as required or be excluded.In addition, although show the particular order of step, this order is only illustrative, and when not departing from the scope of embodiment herein, can utilize any layout suitably of step.

Therefore, technology described herein is that the optimum that E-TREE service provides in EVPN forwards.More specifically, technology herein provide the ingress filtering of known unicast flow and entrance copy situation under the ingress filtering of many destinations flow (clean culture of broadcast, multicast and the unknown).In addition, technology herein provides the optimal utilization of the EVPN network bandwidth, and general simple mode is done like this on EVPN PE, such as, supports the feature had in the dataplane in the current PE of minimum increment formula exploitation.

Although illustrate and described the illustrative embodiment providing in EVPN and there is the optimum E-TREE service forwarded, will be appreciated that can make various other in the spirit and scope of this paper embodiment adapts to and amendment.Such as, embodiment has illustrate and described the embodiment relevant to specific protocol and UNC.But the embodiment more in broad sense is not limited, and in fact can use together with the suitable agreement of other type and/or UNC.Such as, can be benefited from technology herein with E-TREE, other agreement of having multiple spot EVC, L2VPN etc. of root similar.

Aforesaid description is for specific embodiment.But, it is evident that, other change and amendment can be made to the embodiment described, to obtain their some or all advantages.Such as, clear be contemplated that assembly described herein and/or element may be implemented as and be stored in tangible (non-transient state) computer-readable medium (such as, disk/CD/RAM/EEPROM etc.) on software, wherein this software has the program command performed in computer, hardware, firmware or their combination.Therefore be described by means of only example and be not the scope of in addition restriction embodiment herein.Therefore, the object of appended claim is, covers all such changes in the real spirit and scope of embodiment herein and amendment.

Claims

1. a method, comprising:

The frame going to destination medium access control (MAC) address is received at edge (PE) the equipment place of entrance provider in computer network territory;

Judge that described frame is received on root entry ethernet segment or leaf entrance ethernet segment;

Judge that described destination MAC Address is located via root ethernet segment or leaf ethernet segment; And

Be root or leaf respectively based on entrance ethernet segment and destination MAC Address ethernet segment, abandon or forward described frame.

2. the method for claim 1, also comprises:

In the new MAC Address of described entry PE equipment place study; And

Judge that described MAC Address is learnt on root ethernet segment or leaf ethernet segment respectively.

3. method as claimed in claim 2, also comprises:

Based on the mark in MAC advertising line, judge that ethernet segment is root ethernet segment or leaf ethernet segment.

4. method as claimed in claim 2, also comprises:

Based on Border Gateway Protocol (BGP) attribute field, judge that ethernet segment is root ethernet segment or leaf ethernet segment.

5. method as claimed in claim 4, also comprises:

In transmitting, the instruction that ethernet segment is root ethernet segment or leaf ethernet segment is set for each MAC Address.

6. the method for claim 1, wherein described computer network territory is E-TREE, and wherein root communicates with leaf with other roots all, and leaf only communicates with root.

7. the method for claim 1, wherein described frame goes to multiple destination, and described method also comprises:

When described frame is received on leaf ethernet segment, stop any outlet PE equipment being forwarded to by described frame and being marked as leaf.

8. method as claimed in claim 7, wherein, outlet PE equipment have root ethernet segment, leaf ethernet segment or they both, and wherein specific outlet PE equipment has both root ethernet segment and leaf ethernet segment, and described specific outlet PE equipment is marked as root.

9. a device, comprising:

One or more network interface, this one or more network interface communicates as provider edge (PE) equipment in computer network territory, described network interface definition ethernet segment;

Processor, this processor is coupled to described network interface and is configured to perform one or more process; And

Memory, this memory is configured to store the process that can be performed by described processor, and described process can operate when being performed and perform following process:

The frame going to destination medium access control (MAC) address is received at entry PE equipment place;

Judge that described frame is received in root entry section or leaf entrance;

Judge that described destination MAC Address is located via root segment or leaf section; And

Be root or leaf respectively based on entrance and destination MAC Address section, abandon or forward described frame.

10. device as claimed in claim 9, wherein, described process can operate when being performed and perform following process:

Entry PE equipment as described computer network territory learns new MAC Address in this locality; And

Judge that described MAC Address is learnt on root segment or leaf section respectively.

11. devices as claimed in claim 10, wherein, described process can operate when being performed and perform following process:

Entry PE equipment as described computer network territory learns new MAC Address in this locality;

Judge that described MAC Address is learnt on root segment or leaf section respectively;

Based on the mark in MAC advertising line, judge that section is root segment or leaf section.

12. devices as claimed in claim 10, wherein, described process can operate when being performed and perform following process:

Based on Border Gateway Protocol (BGP) attribute field, judge that section is root segment or leaf section.

13. devices as claimed in claim 10, wherein, described process can operate when being performed and perform following process:

The instruction of root segment or leaf section for each MAC Address section of setting in transmitting.

14. devices as claimed in claim 9, wherein, described computer network territory is E-TREE, and wherein root communicates with leaf with other roots all, and leaf only communicates with root.

15. devices as claimed in claim 9, wherein, described frame goes to multiple destination, and wherein said process can operate when being performed and performs following process:

When described frame is received in leaf section, stop any outlet PE equipment being forwarded to by described frame and being marked as leaf.

16. devices as claimed in claim 15, wherein, outlet PE equipment have root segment, leaf section or they both, and wherein specific outlet PE equipment has both root segment and leaf section, and described specific outlet PE is marked as root.

17. 1 kinds of tangible non-transitory computer-readable medium, this computer-readable medium has the software be coded on it, and described software can operate when being executed by processor and perform following process:

Judge that described frame is received in root Ethernet port section or leaf Ethernet port section;

18. computer-readable mediums as claimed in claim 17, wherein, described software can operate when being performed and perform following process:

In the new MAC Address of described PE equipment place study; And

19. computer-readable mediums as claimed in claim 18, wherein, described software can operate when being performed and perform following process:

20. computer-readable mediums as claimed in claim 18, wherein, described software can operate when being performed and perform following process:

21. computer-readable mediums as claimed in claim 18, wherein, described software can operate when being performed and perform following process:

22. computer-readable mediums as claimed in claim 17, wherein, described computer network territory is E-TREE, and wherein root communicates with leaf with other roots all, and leaf only communicates with root.

23. computer-readable mediums as claimed in claim 17, wherein, described frame goes to multiple destination, and wherein said software can operate when being performed and performs following process:

24. computer-readable mediums as claimed in claim 23, wherein, outlet PE equipment have root ethernet segment, leaf ethernet segment or they both, and wherein specific outlet PE equipment has both root ethernet segment and leaf ethernet segment, and described specific outlet PE equipment is marked as root.